A scenario mining method and statistical analysis of driving behavior characteristics of vehicle negotiating a curve based on a data-driven approach

This study delves into comprehending driving behavior traits and augmenting the human-like driving capability of autonomous vehicles by examining driver behavior characteristics during vehicle negotiating a curve using a natural driving dataset. The research establishes a data acquisition system dedicated to capturing driving behaviors, enabling the extraction of multi-dimensional features from natural driving scenarios. By leveraging a significant repository of authentic driving data soured from a diverse pool of driving styles, the study extracts scenarios involving vehicle negotiating curves to conduct statistical analysis and extract driving behavior characteristics within this scenario. The analysis primarily focuses on the comprehensive speed planning of vehicles negotiating a curve, encompassing the deceleration upon entering the curve and acceleration upon exiting it, correlating these actions with the curve features. Ultimately, this research is geared towards furnishing empirical backing for designing automated driving or personalized driving assistance systems that emulate human-like operational attributes.